Olive (Olea europaea L. cv. Picual and Arbequina) trees were grown in the experimental orchard of Instituto de la Grasa, Seville (Spain), with drip irrigation and fertirrigation from the time of complete flowering to fruit ripening.

Four olive branches carrying 100 olive fruit at 28 WAF (turning stage) each one, were collected from different olive trees located in the same field, and transferred to growth chamber where they were incubated at 25°C with a 12 h light/12 h dark cycle, with a light intensity of 300 μmol m⁻² s⁻¹. These incubation parameters attempted to simulate physiological conditions of the tree, and were considered the standard conditions. No significant alterations in the fatty acid composition or SAD genes expression levels were observed in the mesocarp tissue when olive fruits were incubated under the above mentioned standard conditions (Supplementary Figure S2). For stress treatments, a new set of four olive branches with the same characteristics as previously mentioned were collected from the different olive trees and transferred to growth chamber, where the described standard conditions were altered to achieve the stress conditions to be studied. To examine the effect of temperature, the olive branches containing the fruits were incubated at 15°C for the low temperature experiment and 35°C for the high temperature one, at the same standard light intensity. For darkness treatment, light was turned off maintaining the same standard temperature. To study the effect of wounding, the whole surface of the olive fruit was mechanically damaged at zero time exerting pressure using tweezers with serrated tips, so that mesocarp tissue was affected. Each experiment corresponding to a different treatment was carried out at a different day of the same week, to ensure that the olive fruits of the different experiments were in the same stage (28 WAF). Zero time was designated 2 h after the beginning of the light period in every experiment, in order to maintain the natural photoperiod day/night of the olive fruit. At the indicated times, 10 olive fruits were taken from each olive branch, then 1–2 g of olive mesocarp was collected from 5 different olive fruits for RNA isolation, and 1.5 g was collected from the other 5 different olive fruits for lipid analysis. Therefore, in each experiment we used approximately 60 olive fruits from three olive branches, one branch for each biological replicate. In all experiments we incubated four branches, so that in case that we had a problem with a branch, to ensure we had another branch with olives and we could continue with the experiment. Olive mesocarp samples were frozen in liquid nitrogen and stored at −80°C.

Total RNA isolation was carried out according to Hernández et al. (2005) using 1–2 g of frozen olive fruit mesocarp tissue collected from at least five different olive fruit per each of the three biological replicates. Briefly, the frozen olive mesocarp was ground in a pre-cooled mortar with liquid nitrogen and homogenized with the extraction buffer, containing Tris-HCl, NaCl, Na₂EDTA and SDS, and 2-mercaptoethanol. Afterward, nucleic acids were extracted with phenol/chloroform twice, and precipitated with NaAc and ethanol. The nucleic acid pellet was resuspended in DEPC treated water and LiCl was added to precipitate the RNA. The pellet was washed twice with 70% ethanol and resuspended in 25 μl DEPC-water. RNA quality verification, removal of contaminating DNA and cDNA synthesis were performed as described by Hernández et al. (2009), using the TURBO DNA-free kit (Ambion, United States) and SuperScript III First-Strand Synthesis System (Invitrogen, Carlsbad, CA, United States) with oligo (dT)₂₀.

Gene expression analysis was performed by quantitative real time PCR (qRT-PCR) using a CFX Connect real-time PCR System and iTaq Universal SYBR Green Supermix (Bio-Rad, Hercules, CA, United States). Primers for gene-specific amplification were described by Parvini et al. (2016) and are shown in Supplementary Table S1. Reaction mix (10 μL per well) contained 1X iTaq-QPCR Master Mix, 100 nM forward and reverse primers, and 2 μL of cDNA of appropriate dilution, which was selected according to the primers amplification efficiency. The thermal cycling conditions included an initial denaturation step of 95°C for 10 min, followed by 40 cycles of 95°C for 30 s, 60°C for 1 min, and 72°C for 30 s. The melting reaction from 55°C through 95°C, at 0.1°C s⁻¹, following the final step of the PCR, was used to examine the specificity of the PCR amplification and the presence of primer dimers. Additionally, the purity of the PCR products were also checked by agarose gel electrophoresis. PCR efficiencies (E) of all primers were calculated using dilution curves with eight dilution points, twofold dilution, and the equation E = [10^(−1/slope)] – 1. For normalization of the data, the olive ubiquitin2 gene (OeUBQ2, AF429430) was used as an endogenous reference. The qRT-PCR data were calibrated relative to the corresponding gene expression level at zero time for each treatment and cultivar, following the relative quantification by the 2^−ΔΔCt method (Livak and Schmittgen, 2001). The data are presented as means ± SD of the three biological replicates, each having two technical replicates per 96 well plate.

For lipid analysis, 1.5 g of frozen olive mesocarp tissue collected from at least five different olive fruits per each of the three biological replicate, were used. Olive fruit mesocarp tissue was firstly treated with isopropanol at 70°C for 30 min to inactivate endogenous lipase activity. Extraction of lipids was carried out according to Hara and Radin (1978), followed by their separation by thin layer chromatography as described by Hernández et al. (2008). Acid-catalyzed transmethylation of the different lipid preparations was performed to obtain the corresponding fatty acid methyl esters (Garcés and Mancha, 1993), which were analyzed by gas chromatography (Román et al., 2012). The internal standard used to calculate the lipid and fatty acid content in the samples was heptadecanoic acid. Results are expressed in μg of the sum of UFA per mg of FW, and are presented as means ± SD of three biological replicates.

To investigate how temperature regulates UFA synthesis in olive mesocarp, olive branches from Picual and Arbequina cultivars holding olive fruits at 28 WAF, which correspond to turning stage, were incubated at low (15°C) and high (35°C) temperature with a 12 h light/12 h dark cycle, for 24 h. Lipid and SAD genes expression analysis were carried out using olive mesocarp at different times after incubation at the aforementioned conditions.

The effect of low temperature (15°C) on the UFA content and SAD transcript levels was different in the two cultivars studied. Low temperatures reduced the UFA content at the beginning of incubation in both cultivars, and then recovered practically to the initial levels at 6 h of incubation. These levels were maintained in Arbequina mesocarp until the end of the experiment. However, in Picual mesocarp the UFA levels increased during the 24 h period, reaching about 20% more than at the beginning of the treatment (Figure 1A). These results correlated well with the expression levels of SAD genes in mesocarp tissue exposed to the cold treatment. In this way, while Arbequina SAD transcripts remained practically constant throughout the treatment, in Picual mesocarp the expression levels of the three SAD genes showed a transient and significant increase during the first 3 h of incubation, with SAD1 undergoing a 10-fold increase in gene expression compared to SAD2 and SAD3 with 6- and 3-fold increase, respectively (Figure 1B). The effect of incubation at 15°C observed in Picual mesocarp is the general response to low temperature changes for desaturases, since there are numerous reports describing that low temperature causes an increase in the UFA content (Los and Murata, 1998). With only two exceptions in lima bean (Zhang et al., 2011) and avocado (García-Rojas et al., 2012), an induction of SAD genes by cold stress has been reported in different plant species. For instance, Wang et al. (2013) described that in Ginkgo biloba leaves, SAD mRNA levels increased transiently, reaching a maximum at 6 h after incubation at 4 or 15°C, and then decreased after 24 h, analogous to what we observed in the Picual olive mesocarp. Similarly, a transient induction of SAD gene was observed in tea leaves incubated at 4 or −5°C for 24 h (Ding et al., 2016). In a like manner, an increase in SAD transcript in response to low temperature has been described in avocado fruit (Madi et al., 2003), potato leaves (Vega et al., 2004; De Palma et al., 2008), rape hypocotyl (Tasseva et al., 2004), and soybean seed (Byfield and Upchurch, 2007), which was also accompanied by an increment in the UFA content. The role of SAD gene in plant cold stress has also been demonstrated in transplastomic tobacco plants expressing a wild potato SAD gene, that exhibited increased UFA content and improved cold tolerance (Craig et al., 2008). Furthermore, the rice cold inducible transcription factor Osmyb4, which is involved in the cold stress response, has been shown to transactivate the wild potato SAD gene promoter (Vannini et al., 2004).

To gain further insight about the cold stress response observed in olive mesocarp from Picual cultivar, we decided to analyze the UFA content in the different lipids classes from this tissue incubated at 15°C for 24 h, to investigate whether the detected increase in SAD genes expression leads to an increase in the UFA content in membrane lipids. We observed that the UFA content in triacylglycerol decreased throughout the incubation at 15°C, while UFA levels increased significantly in the membrane phospholipids (Table 1). In addition, we observed an increase in all of the UFA, being oleic acid the one that increases the most (Supplementary Table S2). These data suggested that low temperature incubation induced a mobilization of UFA from storage lipids into microsomal membrane lipids.

We have previously reported an induction of FAD2 genes in mesocarp tissue incubated at 15°C for 24 h, although non-significant differences were detected in the linoleic acid content of microsomal and plastidial membranes (Hernández et al., 2011). In that study, we suggested that the timescale could be too short to observe effects, although the existence of post-transcriptional regulatory mechanism could not be discarded. In this work (Supplementary Table S2), we noticed that the increase in oleic acid in membrane phospholipids took place from the beginning of the cold treatment (1 h after incubation), whereas a very slight increase of linoleic acid was detected in phosphatidic acid, phosphatidylserine, phosphatidylcholine, and phosphatidylethanolamine at longer times (6–24 h after incubation), suggesting that low temperature induced the synthesis of oleic acid in the short–term, while the increase of linoleic acid occurs at longer incubations periods. These results indicate that the observed changes in the transcript levels of SAD genes caused by low temperature are accompanied by the adjustment of UFA content of microsomal membrane lipids and modulation of membrane fluidity in olive mesocarp cv. Picual.

The fact that low temperatures did not increase either SAD expression levels or the UFA levels in Arbequina mesocarp could be related with a lower cold tolerance of this cultivar with respect to Picual. In this sense, Vega et al. (2004) reported that the expression of SAD gene increased during cold acclimation only in Solanum commersonii, a species capable of cold acclimation, and not in the cultivated non-cold acclimating Solanum tuberosum species, although the latter had a greater amount of constitutive SAD gene expression, indicating that the changes in transcript accumulation observed in S. commersonii may be related to its capacity to cold acclimate. However, these data contrast slightly with those found by De Palma et al. (2008), which provided evidence that the freezing tolerant S. commersonii plants have a higher constitutive transcript levels of SAD gene. Nevertheless, these results altogether suggest that SAD transcript accumulation plays a key role in cold tolerance. This assumption was further confirmed by Li et al. (2015), who reported that SAD overexpression caused an increase in membrane linoleic acid content, which improved the cold acclimation capacity of transgenic potato plants.

With respect to high temperature, when we incubated the olive branches with turning olive fruits at 35°C for 24 h, we observed a decreased in the UFA content 1 h after the incubation, to subsequently recover to the initial values throughout the incubation period in both cultivars, reaching even higher values than the initial ones after 24 h of incubation (Figure 2A). The changes observed in UFA content are mainly detected in TAG for both cultivars (Supplementary Tables S3, S4). These modifications in UFA levels did not correlate well with the SAD genes expression pattern detected in mesocarp tissue exposed to high temperature. We observed in Figure 2B that the expression levels of the three SAD genes decreased during the incubation at 35°C, although SAD genes expression patterns were different in both cultivars. Specifically, in cv. Picual, SAD1 and SAD2 transcripts slightly increased at 0.5 and 1 h of incubation, the three genes expression levels were similar to initial values at 3 h, and then decreased considerably after 24 h of incubation. However, in cv. Arbequina the three SAD transcripts decreased from the beginning of the treatment. The downregulation of SAD genes has been previously reported by Wang et al. (2013) in leaves of ginkgo grown at 35 or 45°C, and by Lu et al. (2013) in Pinellia ternata leaves incubated at 35°C for 24 h. In addition, when soybean plants were incubated under warm conditions, the SAD-A and SAD-B genes expression levels decreased in the seeds, but with negligible effect on the seed stearate content (Byfield and Upchurch, 2007). The lack of correlation between the effect of high temperature on SAD genes expression levels and the UFA content could be explained by several factors. In particular, the high temperature regulation of oleic acid synthesis may be mediated by post-transcriptional mechanism. Several cases of temperature-related post-transcriptional mechanisms have been reported for the oleate desaturases. In sunflower seeds, it has been described that changes in temperature bring about shifts in the very low endogenous oxygen concentration, which affect FAD2 activity reversibly, without having an effect on gene transcription (Rolletschek et al., 2007). Besides, Tang et al. (2005) identified two domains that appear to be important in mediating the temperature-dependent instability of the soybean FAD2-1A isoform when expressed in yeast. However, to date, no mechanism of post-transcriptional regulation by temperature has been described for SAD enzymes.

To test whether darkness alters the UFA content and SAD transcript levels in olive mesocarp, olive branches from Picual and Arbequina cultivars holding olive fruits at 28 WAF (turning stage) were incubated at 25°C in the darkness, for 24 h. Although we did not observe any significant difference in the total UFA content during the incubation period, a decrease in the three SAD genes expression levels was detected in both cultivars (Figure 3). SAD1 and SAD3 transcripts levels decreased considerably in both cultivars from the beginning of the treatment, so that after 24 h of incubation reached a reduction of 90-fold in comparison to the initial levels. However, the decrease in SAD2 expression levels was about 50-fold after 3 h of incubation, maintaining these levels in Picual cultivar and recovering the initial values in the case of Arbequina (Figure 3B).

Several evidences indicate that light may play a regulatory role for plant desaturases, although data are scarce in this area. A light-dependent transcriptional regulation of oleate desaturase genes has been described in olive mesocarp (Hernández et al., 2011), similarly to that reported by Kargiotidou et al. (2008) in cotton cotyledons, where FAD2 expression was reduced after a 24 h incubation in the dark. In the same way, FAD7 and FAD8 are also transcriptionally regulated by light in different plants (Nishiuchi et al., 1995; Horiguchi et al., 1996; Collados et al., 2006). In general, it appears that increasing light conditions are associated with an enhancement in desaturases genes expression, with the converse also holding true in darkness conditions. Although this is the first time that the effect of darkness on SAD genes expression levels is reported in plants, the regulatory role of light on Δ9 desaturase genes has been studied before in other photosynthetic organism. Kis et al. (1998) described that in the cyanobacteria Synechocystis PCC 6803 desaturases genes were strongly induced by light, except for the Δ9 desaturase, that was not significantly affected. On the other hand, Ma et al. (2018) observed that the expression levels of SAD gene from the green microalgae Haematococcus pluvialis were significantly upregulated by light, which correlated well with an increase in oleic acid. These results are in agreement with the decrease in SAD genes expression levels observed in olive mesocarp in darkness conditions, although we did not observe a decrease in the total UFA content.

To further investigate the effect of darkness in oleic acid synthesis, we decided to analyze the UFA content of the different lipid classes in olive mesocarp subjected to darkness conditions, in order to elucidate whether the decrease in the expression levels of SAD genes affects the UFA content of a specific lipid. Interestingly, we only observed a decrease in the UFA content of the galactolipids, MGDG and DGDG, in both cultivars (Figure 4), although with different behavior. While in Picual mesocarp, the UFA content decreased considerably during the 24 h of incubation under darkness conditions, in Arbequina mesocarp, the UFA levels were reduced after 1 h of incubation, and then recovered, almost reaching the levels detected at the beginning of the treatment. This effect of the dark on the pattern of UFA content in galactolipids correlated quite well with that of SAD2 expression levels in both cultivars and not with SAD1 and SAD3, which suggests, that the reduction of SAD2 transcript due to the dark incubation could be the responsible for the decrease in UFA in galactolipids. In fact, we noticed that the decrease detected in the UFA content of MGDG and DGDG during the dark incubation of olive mesocarp is mainly due to a reduction in oleic acid content (Table 2). Therefore, we can conclude that the down-regulation of SAD2 gene during the dark incubation is the responsible for the reduction in the UFA content in the main chloroplast lipids, MGDG and DGDG. Furthermore, this light-dependent regulation detected in olive mesocarp is cultivar-dependent, since the restoration of SAD2 gene expression levels and UFA content in galactolipids after 24 h of incubation observed in Arbequina cultivar, was not detected in Picual mesocarp. The fact that the effect of darkness was noticeable on chloroplast-localized lipids is not unexpected. Gemmrich (1982) observed in Ricinus communis cultures, that the light-induced changes in lipid composition were associated with thylakoid formation. Strong light induces ultrastructural changes in chloroplasts, so that the area of thylakoid system on chloroplast sections increases, there is an accumulation of chloroplast-localized lipids, like MGDG, and the unsaturation index of fatty acids is elevated, being the relative content of linolenic acid the one that increases the most (Kislyuk et al., 2013).

The involvement of a Δ9 stearate desaturase in defense mechanisms was firstly demonstrated by Xing and Chin (2000), who reported that the expression of a yeast Δ9 desaturase in eggplant enhanced its resistance to Verticillium dahliae. To investigate possible changes in the UFA composition and SAD genes expression of olive mesocarp in response to wounding, olive branches from Picual and Arbequina cultivars holding olive fruits at 28 WAF (turning stage) were incubated using standard conditions except that olive fruit were mechanically damaged at zero time. A slight increase in the UFA content was observed when lipid analysis was performed at different times of incubation after wounding, although with some cultivar differences. While in cv. Picual the increase in UFA content after wounding was slow and progressive, in Arbequina mesocarp the UFA content increases rapidly after 1 h of incubation, continues rising until 6 h and then, the levels were maintained until the end of the incubation (Figure 5A).

In contrast, except in the case of Picual SAD3 gene, olive SAD genes transcript levels decreased after wounding in both cultivars (Figure 5B). It has been demonstrated that a mutation in the Arabidopsis gene ssi2/fab2, which encodes a SAD, resulted in the reduction of oleic acid levels, causing the constitutive defense response in the mutant plant (Kachroo and Kachroo, 2009; Xia et al., 2009). Analogous results were observed in OsSSI2-knockdown plants in rice (Jiang et al., 2009), and in GmSAD-silenced soybean (Kachroo et al., 2008). In addition, overexpression of the TaSSI2 in ssi2 Arabidopsis mutant plants resulted in restoration of oleic acid and, thereby, rescued other ssi2-associated phenotypes (Song et al., 2013). Further studies revealed that the reduced oleic acid levels triggered the transcriptional up-regulation of pathogenesis-related genes, the genes governing synthesis of salicylic acid, and nitric oxide responsive nuclear genes, thus activating disease resistance (Venugopal et al., 2009; Mandal et al., 2012). In this sense, we also detected in olive mesocarp a discrete reduction of 1.24% in Picual and 2.88% in Arbequina in oleic acid levels after wounding (Table 3). Since olive SAD2 has been reported to be the main gene contributing to the oleic acid content in olive mesocarp (Parvini et al., 2016), the down-regulation of olive SAD1 and SAD2 genes observed in response to wounding could be responsible for the decrease in the oleic acid proportion and, consequently, trigger the defense response.

Unlike SAD1 and SAD2 genes, SAD3 increased its transcript levels transiently after 1 h in cv. Picual, and returned to initial values at 24 h of treatment (Figure 5B). Remarkably, SAD3 gene has been previously shown to be induced in olive leaves infected by Spilocaea oleagina (Benítez et al., 2005), and Verticillium dahliae and Fusarium spp. (Trabelsi et al., 2017). Similar results to olive SAD3 gene induction were observed in yellow lupine, avocado and tea. Zaborowska et al. (2002) reported an increase in SAD gene transcripts in yellow lupine nodules from 12 days after infection with Bradyrhizobium sp. (Lupinus). In the same way, SAD expression levels increased in response to wounding in avocado fruit (Madi et al., 2003) and tea leaves (Ding et al., 2016). Interestingly, we observed that the reduction in the oleic acid levels was accompanied by an increase in palmitolinoleic and linoleic acids at 24 h after wounding in both cultivars (Table 3). We have reported before this increase in dienoic fatty acids induced by wounding, showing that FAD2 genes are involved in the wounding response of olive fruit mesocarp, and causing an increase in the content of palmitolinoleic and linoleic acids in microsomal lipids (Hernández et al., 2011). In the same study, we also suggested that the synthesis of palmitolinoleic acid is a consequence of the simultaneous induction of SAD and FAD2 genes in olive fruit mesocarp in response to wounding. Since, in this work, we have observed a specific induction of the SAD3 gene (Figure 5B), it is tempting to speculate that the SAD3 isoform is the one involved in this response mechanism. These dienoic fatty acids probably serve as precursors of a different set of oxylipins involved in plant defense, generated by the lipoxygenase pathway (Weber, 2002).